Department of Statistics Unitmark
Dietrich College of Humanities and Social Sciences

In honor of the Department of Statistics' 50th Anniversary, we are highlighting our outstanding faculty and alumni of the past half-century.

Chris Paciorek

Carnegie Mellon Statistics Department alumnus Chris Paciorek (Ph.D. 03) has four different titles in the Statistics Department at the University of California, Berkeley.  “First, I have a research position in which I bring in grant money to support my work on statistical applications in ecology, climate, and public health and on software development for hierarchical modeling,” he said.  “Second, I teach one class a year - for the last five years this was a graduate-level statistical computing class, but this year I am teaching a Bayesian statistics class.  “Third and fourth, I have a staff position in which I provide computing support to researchers in statistics and economics, and more broadly on campus through the Berkeley Research Computing program.  “I like that I can work on big problems in my research but when the research is bogged down or seems only indirectly related to solving real-world problems, I still have the small successes of helping solve others' computing problems,” he said. 

Chris grew up outside of Boston, and after undergraduate work in a variety of locations, enrolled as a graduate student in ecology at Duke University.     “When I started taking statistics classes, I realized quickly that I liked it, particularly the more creative aspects of developing models for applications, and the elegance of the Bayesian perspective,” he said.  A year later, he applied to the Statistics Ph.D. program at Carnegie Mellon.  “The breadth of topics and exposure to both the Bayesian and frequentist perspectives prepared me to work on a broad array of problems from both perspectives, which I've since done.  “Valerie Ventura taught the masters-level statistical theory class, and that gave me a very solid foundation for later classes and the rest of my career.  “The Advanced Data Analysis (ADA) project and class were a great in-depth immersion into statistical applications and led directly to my work with a climate scientist, James Risbey, the outside collaborator on my ADA and PhD thesis.  “I'm currently working with a post-doc here at Berkeley on Bayesian multiple testing in a climate application. It turns out that work relates directly to a side project I did while at CMU, working with James and Valerie on use of the false discovery rate in climate applications,” he said.  “Chris is great!” said Professor Ventura.  “He always does the right thing, from cleaning messes in the statistics lounge, raising orphan bats by hand, teaching math to prison inmates to help rehabilitate them, to taking to the streets to protest some unfair policy, and occasionally being tackled by the FBI for showing a tad too much fervor!  “But he is not always serious and thoughtful; he sometimes does funny things, for example, using a totally crooked broken curtain rod found at the scene of his PhD defense to replace a forgotten pointer.  “Chris was a great student, too. A particularly nice memory was his taking a PhD seminar I taught to catch up on the latest MCMC advances.   Seminars can be a flop if no one participates, but Chris made it great fun, coming prepared, opinionated, and willing to hold his own against me.  “Advising him on his ADA was also a breeze: he instantly took ownership of his project, proving to be independent, imaginative, and industrious.  “ Our second paper together is one of my favorites. The statistical method was not particularly novel, but bringing it to climatology with the right tone and application certainly was.  “My computer has pinged me regularly since then when it gets cited.  “But perhaps the best of it was the most positive and funny review we got from a referee who wholly disagreed with the premise of the paper,” she said.  

Chris’ Ph.D. research focused on nonstationary covariance structures for Gaussian process models with application to climatological and other spatial data and to nonparametric regression modelling.  “I'm certainly pleased that Chris chose me as a thesis advisor,” Professor Mark Schervish said.  “I believe that I learned as much or more from working with him than he learned from working with me.  “He did some excellent work for which I wish I could take some credit, but he really knew what he was doing and how to get it done,” he said.  “Mark provided savvy feedback on my ideas and, at key moments, pointed me in very productive directions critical to the success of the thesis; plus, his understated advising style worked perfectly for me as an older and independent student,” Chris said. 

Today, he calls his collaboration with a small group to develop a software platform called NIMBLE, which functions in some ways like JAGS and BUGS in terms of providing automated MCMC for arbitrary hierarchical models, his “most exciting” statistical research.  “What's new is that it also provides much more flexibility in terms of controlling the behavior of the MCMC (choosing samplers, blocking parameters) and allowing users to program their own algorithms (e.g., sequential Monte Carlo and MCEM) that can be applied to models specified in the BUGS language.  “I've also done a lot of work on a project called PalEON that seeks to use paleoecological and paleoclimate data to calibrate and assess ecosystem models that are used to analyze carbon storage in forests and understand ecosystem feedbacks on climate change.  “Much of this has involved calibrating statistical models that relate paleo proxy data to ecological and climate data and using the models to estimate ecological and climate variables over the past few thousand years.   “One of my collaborators describes this as building a 'time machine', but as people who are fascinated by natural landscapes, really what we'd both rather do is actually be able to go back and look at the forests and prairies before European settlement,” he said.  As for the future, Chris said he is “pretty settled” in Berkeley, and with a native Californian girlfriend adding to his investment in the region.  “After seven years here, I've somewhat lost my sense that California and the West are strange, exotic places, but not entirely,” he said.